Everything you need to master the AI tools.

Overview

The DevToolAI AI Toolkit is a collection of six focused developer tools powered by Google Gemini. Instead of one giant chatbot, each tool is prompt-engineered for a specific job: converting code between languages, explaining unfamiliar snippets, debugging bugs, generating unit tests, writing READMEs and producing SQL from natural language.

Every tool runs against the same underlying model but with a different system instruction, temperature and post-processing rule. That is what makes the output feel like a purpose-built utility rather than a general purpose LLM chat.

All AI processing happens on our servers via a secure server-side call to the Gemini API. No API keys are ever exposed to the browser, and your code is not persisted — we only log anonymised metadata (tool name, input length, latency) so we can improve the service.

Quick start

Get an answer in 15 seconds:

1. Open any AI tool from the AI Tools landing page.
2. Click one of the Try: chips at the top to load a real example.
3. Or paste your own input into the left panel. The character counter turns amber above 17,000 characters and red at the 20,000 hard limit.
4. Hit Generate. The answer streams token-by-token into the right panel.
5. When you are done, use Copy to grab the output or Download to save it as a file.

Tool reference

The full reference for every tool below — what it does, expected input format, and tips for best results.

AI Code Converter

Translate code between programming languages instantly.

Convert code between JavaScript, TypeScript, Python, Java, C#, Go, Rust, PHP, Ruby, Kotlin, Swift and more — preserving logic, idioms and readability.

Expected input: Source code (mention source & target language on the first line, e.g. "// Convert from Python to TypeScript")

Tips for best results:

• Put the source and target language in a comment on the first line — e.g. `// Convert from Python to TypeScript`.
• For framework migrations (React → Vue), name both frameworks explicitly.
• Include imports if you want the output to include them too.

Open Code Converter →

AI Code Explainer

Understand any code — line by line, in plain English.

Paste any snippet and get a clear, senior-engineer-style walkthrough — purpose, step-by-step logic, complexity, and edge cases.

Expected input: Code to explain

Tips for best results:

• Paste a focused function rather than an entire file for a tighter explanation.
• Add a comment saying "Focus on X" if you want the walkthrough to zoom in on a specific part.
• Beautify minified code before pasting for better line-by-line breakdowns.

Open Code Explainer →

AI Code Debugger

Find bugs, understand root cause, get a fixed version.

Paste buggy code (plus the error/stack trace if you have one) and Gemini will list every bug, explain the root cause, and return a corrected version.

Expected input: Buggy code + optional error message

Tips for best results:

• Always include the error message or stack trace when you have one.
• If the bug is subtle, describe the expected vs actual behaviour in a comment.
• For multi-file bugs, paste just the relevant snippets and note each filename in a comment.

Open Code Debugger →

AI Unit Test Generator

Auto-generate tests for functions, classes and modules.

Generate comprehensive unit tests (happy paths, edge cases and errors) in the framework idiomatic to your language — Jest, pytest, JUnit, Go testing and more.

Expected input: Function or module to test

Tips for best results:

• Mention your framework on the first line (`// Framework: Vitest`) to override the default.
• Include the function signature and types — the more the AI knows about inputs, the sharper the tests.
• Add a comment listing any known edge cases you want covered.

Open Unit Test Generator →

AI README Generator

Turn project details into a polished README.md.

Describe your project (name, stack, features, scripts) and get a complete GitHub-ready README with badges, TOC, installation, usage, and contributing sections.

Expected input: Project details

Tips for best results:

• Structure your input as key-value pairs (Project, Description, Stack, Features, Scripts).
• Mention your audience ("Audience: beginner developers") to tune the tone.
• Include repo URL and license so placeholders come out closer to production-ready.

Open README Generator →

AI SQL Query Generator

Describe what you want — get correct SQL.

Provide your schema (or describe your tables) plus a natural-language question, and Gemini will produce the SQL query with an explanation.

Expected input: Schema + question

Tips for best results:

• Paste the DDL or list your tables with columns — the AI is only as good as the schema you give it.
• Name the SQL dialect at the top: Postgres, MySQL, SQL Server, SQLite, BigQuery, Snowflake.
• For analytical queries, mention window functions or CTEs explicitly if you want them used.

Open SQL Generator →

Best practices

• Be specific about your target. For Code Converter, put a comment on the first line like // Convert from Python to TypeScript. For SQL Generator, mention the dialect (Postgres, MySQL, BigQuery) at the top.
• Provide context, not just code. For Debugger, paste the error/stack trace alongside the code. For Test Generator, mention your framework (Jest, Vitest, pytest).
• Keep inputs focused. Paste one function or a small module rather than an entire file. Focused inputs get focused answers.
• Iterate. If the first answer is too generic, add a comment explaining what you specifically want to know or the constraints of your codebase.
• Always review AI output. Gemini is excellent but occasionally hallucinates. Treat every generation as a first draft you sanity-check.

Privacy & security

Your privacy is the priority. Here is exactly what happens with your input:

• End-to-end HTTPS. All requests are encrypted in transit.
• Server-side only. The Gemini API key never leaves our servers. It is never exposed to your browser.
• No storage of your input. The code / prompt you paste is forwarded to Gemini for the current request only. We do not persist it.
• Anonymous analytics. We log the tool used, input length, output length and latency for capacity planning and abuse prevention. We do not log the input itself.
• Google's data policy. Once your input reaches Gemini, Google's API terms apply.

Limits & error handling

• Input size. Each tool accepts up to 20,000 characters (roughly 5,000 tokens). The counter above the input turns amber near the limit and red when exceeded.
• Rate limits. If we (or Gemini) are rate-limited you will see: "AI is rate-limited right now. Please try again in a few seconds."
• Authentication errors. Rare, but if the underlying API key is misconfigured you will see: "Gemini authentication failed."
• Timeouts. Very large outputs may time out around the 60 second mark. Retry with a smaller input.
• Empty responses. If Gemini returns an empty response we surface: "Gemini returned an empty response." Usually retrying with more context solves this.

FAQ